I like Nicholas Wade, and think that his latest NY Times piece on basic research is worth reading. However, I take issue with his overly simplistic characterization of how research works:

Basic research, the attempt to understand the fundamental principles of science, is so risky, in fact, that only the federal government is willing to keep pouring money into it. It is a venture that produces far fewer hits than misses....

If basic research is fraught with such a high failure rate, why then does it yield such rich economic returns? The answer is that such government financing agencies as the National Institutes of Health and the National Science Foundation are like the managers of a stock index fund: they buy everything in the market, and the few spectacular winners make up for all the disasters.

This is not right.* Wade goes astray in thinking of science in terms of hits and misses. Basic research is not like being at bat, with the occasional single base hit or home run being the exception in a sea of strikeouts.

Most research is simply conventional and incremental. Most of the time it's not a miss, a disaster, or a failure - it's a small, sometimes not too surprising addition to our knowledge of a subject. Most research projects and NIH grants end in success, not failure - but the successes are usually small.  In fact, there probably aren't enough failures, because, unlike the venture capitalists Wade compares it to, the NIH is very unwilling to take risks in search of the spectacular winner.  Funded projects are the ones almost guaranteed to work.

In any case, you can ask whether our big, spectacular successes are worth all the money we spend on smaller, incremental successes. But it's flat-out wrong to call most research a failure.

*However, the individual researcher can expect weeks or months of failed experiments before finally figuring out how to get things to work. Most scientists fail on a day-to-day basis, but still succeed in the long run.